Rapidus, rapidly running through funds, needs $700M for 2nm chip plant

[-]

Exist50@reddit

In addition to the ¥100 billion being sought from private investors, the Japanese government has promised a considerable pot of money to aid in its endeavor, with the Japan Times reporting in May that the government had ¥920 billion ($6.4 billion) earmarked for the project.

So with government providing 90+% of the funding, realistically this is state owned.

Regardless, I think this whole thing feels like an attempt to milk the government for money with no genuine expectation of ever following through. Skipping straight to a working 2nm fab from literally nothing but some IBM research research is just madness. There's a reason the private sector isn't putting in significant funds.

[-]

Mahalangur@reddit

That's how it is in fab industry. TSMC is dominating the industry and it started out from shitton of government funding in the beginning. In fact, it was the Taiwanese government that reached out to the Morris Chang to be leader of the company that would be funded by the state. In a nutshell, they pretty much gave him a blank check asking if he can run the business.

[-]

Exist50@reddit

This seems to be the reverse. The company is founded then asks the government for money to actually function.

[-]

Legal-Insurance-8291@reddit

Realistically I think we're going to get to a point where it's basically impossible to keep up with government aid. The costs and complexity these days are just too much for a private company to handle.

[-]

Winter_2017@reddit

That was a large portion of the reasoning behind the Rapidus plan. Japan realized that after 2nm any new fab would cost too much to even be government funded, and they wanted capability in their country.

[-]

peakbuttystuff@reddit

It's a jobs program that will yield some.sort.of.fab capacity in 10 years that will go straight into government projects that demand locally sourced microchips.

It's just a huge transfer of government funds to the private sector.

[-]

HeroYouKey_SawAnon@reddit

It’s not a jobs program. Other Japanese fabs have noted a severe labor shortage, especially TSMC who have more or less publicly confirmed that their third Japanese fab will only be constructed once more talent is ready. Like pretty much every global advanced fab project, it is a “national security” national ego thing.

[-]

peakbuttystuff@reddit

National security is the biggest job program usually. It's a buzzword like protecting the children.

Ask any US congressman with LM factories in their state if they want to cut down military expenditure.

[-]

HeroYouKey_SawAnon@reddit

My point is that it’s not a jobs program since the jobs are already in demand from the rest of the industry. If anything it worsens the Japanese silicon job market by stalling out other fabs and forcing Rapidus to take an ‘automation heavy approach’ whatever that means. If the JP gov wanted just long term job support for this domestic industry there are better paths it could have taken but they chose the shiny leading edge option instead.

[-]

peakbuttystuff@reddit

The legions of construction companies love these contracts. I mean it's Japan so probably it's either owned by Sony, Mitsubishi or Toyota, but the biggest winners are the construction companies and their workers.

It's a jobs program. Private companies receiving public money.

[-]

HeroYouKey_SawAnon@reddit

You are doubling down without reading what I said. If you are determined to categorize this as a jobs program no matter what then I can’t convince you but that simply isn’t true. There is already a shortage of labor for fab operation as well as construction. Rapidus is competing for limited construction labor from other fab construction projects as well as general industry and housing construction in the Hokkaido area and nationwide. The government could have funded many more physical fabs over a longer time period from established companies like Sony or Renesas to create longer term jobs, but chose instead to chase the fewer buildings but bigger prestige Rapidus. Toyota finally picked their horse by backing JASM whose third fab is literally delayed largely due to the Rapidus project. This is clearly a prestige and nationalism move not a jobs program.

[-]

Legal-Insurance-8291@reddit

This is no the same as the US where we want more good jobs, but don't actually educate people well enough to do those jobs. I grew up in a poorer state and there were issues even staffing automotive jobs because people literally couldn't read and write.

[-]

HeroYouKey_SawAnon@reddit

FWIW in Japan there is huge investment happening in universities and overseas learning to create more silicon talent. Especially at the Unis of Tokyo and Tohoku. But the graduates from this investment won’t arrive for several more years and Rapidus is gonna hoarde a bunch of them which is why TSMC/JASM fabs are delayed and others like Micron and Sony are wondering publicly what they are gonna do. Also there is a huge wall coming to Japan as the main silicon talent right now are old men gearing up to do their ‘one last project’, what happens when they retire is going to be…interesting.

[-]

kingwhocares@reddit

Only if it was. It's like giving away taxpayers money for free.

[-]

7silverlights@reddit

They probably will not succeed in their promises or at least volume, but what if they do leap frog intel and tsmc? How would the semi industry and even global politics change?

[-]

XiaoPenguin@reddit

This is almost China levels of corporate grifting/corruption. Of course no private investor is putting money into this.

[-]

ledfrisby@reddit

Nobody knew chip fabs could be so complicated! But seriously though, straight to 2nm seems ambitious, doesn't it?

And a bit off-topic, but do you think the author was patting themself on the back for that headline?

[-]

RazingsIsNotHomeNow@reddit

Yeah. Honestly this seems doomed to fail. Even assuming they succeed in producing 2nm like nodes they would be substantially behind from a technology point of view to compete with Tsmc, Samsung and Intel. There's a lot more that goes into each node than simple die shrinks. There's a matter of interconnects, back side power delivery, gate all round, 3D stacking, etc. that all heavily affect the performance of a node. I am not sure who they hired that they believe they can leap frog every other manufacture but they must be smoking something good.

[-]

kingwhocares@reddit

How does a company founded in 2022 start focusing on one of the most advanced nodes in 2 years!

[-]

Exist50@reddit

By not actually caring if they achieve it.